Re: [PATCH v3 1/4] ACPI: Support system notify handler via .sys_notify

[Toshi Kani <toshi.kani@hp.com>]

On Wed, 2012-11-28 at 23:49 +0100, Rafael J. Wysocki wrote:
> On Wednesday, November 28, 2012 03:33:27 PM Toshi Kani wrote:
> > > > > > > > I think it has some challenge as well.  We bind an ACPI driver with
> > > > > > > > device_register(), which calls device_add()-> kobject_add().  So, all
> > > > > > > > non-present ACPI device objects will show up in sysfs, unless we can
> > > > > > > > change the core.  This will change user interface.  There can be quite
> > > > > > > > many non-present devices in ACPI namespace depending on FW
> > > > > > > > implementation.
> > > > > > > 
> > > > > > > If additional devices appear in sysfs, that's not a problem.  If there
> > > > > > > were fewer of them, that would be a real one. :-)
> > > > > > 
> > > > > > I see.  I guess this means that once we expose all non-present devices
> > > > > > in sysfs, we cannot go back to the current way.  So, we need to be very
> > > > > > careful.  Anyway, this model requires separate handling for static ACPI
> > > > > > [1] and dynamic ACPI [2], which may make the state model complicated.
> > > > > > 
> > > > > > 1. Static ACPI - No creation / deletion of acpi_device at hot-plug.
> > > > > > 2. Dynamic ACPI - Create acpi_device at hot-add, delete at hot-remove.
> > > > > 
> > > > > Sure.  The complication here is that we'll need to handle the removal of
> > > > > a bunch of struct acpi_device objects when a whole table goes away.
> > > > > 
> > > > > However, first, we don't seem to handle table unloading now.  At least
> > > > > acpi_unload_parent_table() is not called from anywhere as far as I can
> > > > > say.  Second, we'll need to handle the removal of struct acpi_device objects
> > > > > _anyway_ on table unload, this way or another.
> > > > 
> > > > AML is the one that requests loading/unloading of SSDT for dynamic ACPI.
> > > > In hot-add, _Lxx method executes AML_LOAD_OP or AML_LOAD_TABLE_OP to
> > > > load SSDT and then sends a notification to the OS.  In hot-remove, _EJ0
> > > > method executes AML_UNLOAD_OP to unload SSDT.  Of course, ACPICA does
> > > > the actual work on behalf of AML.  But this is not visible to ACPI core
> > > > or drivers, unless it checks ACPI namespace to see if any device objects
> > > > disappeared after _EJ0.
> > > 
> > > Oh, we have a handler for that event, but we don't really use it. :-)
> > > 
> > > And I wonder what happens to the struct acpi_device objects associated with
> > > the ACPI handles in the table being unloaded?
> > 
> > If we use an ACPI handle that does not associate with a device object,
> > ACPICA returns AE_NOT_FOUND or AE_NOT_EXIST.  But, we should remove
> > acpi_device that does not have its associated ACPI object.  Currently,
> > we create acpi_device on hot-add and remove it on hot-remove, so it is
> > OK.  But if we start creating acpi_device objects for non-present
> > devices, we need to worry about if acpi_device objects indeed have their
> > associated ACPI objects.  That's the complication I mentioned above.
> 
> My question was about something different.  Is it actually guaranteed
> that all struct device objects associated with the given ACPI table will
> be removed before that table is unloaded?

Yes, it is guaranteed currently.  acpi_bus_hot_remove_device() calls
acpi_bus_trim(), which deletes all acpi_device objects under of the
notified object.  It then calls _EJ0 of the notified object, which
unloads all ACPI objects under the notified object. 

Thanks,
-Toshi